www.ijecs.in International Journal Of Engineering And Computer Science ISSN:2319-7242 Volume 2 Issue 6 June, 2013 Page No. 2012-2018

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2012

A NOVEL METHOD TO DETECT BONE CANCER USING IMAGE FUSION AND

EDGE DETECTION

Nisthula P, Mr. Yadhu.R.B.

Calicut University, kerala, India

nisthulanair@gmail.com

Asst.Professor,Calicut university, Kerala , India

Abstract

Employing an efficient processing technique is considered as an essential step to improve the overall visual representation of

clinical images, and as a consequence provides better diagnosis results. This paper employs an easy, fast and reliable

technique to detect cancerous tissue in bone by using different image processing techniques such as contrast enhancement,

edge detection and image fusion. The experimental results show, the proposed method could obtain the smooth image with

edge showing the disease affected part without the spatial and spectral noises.

Key words: contrast enhancement, edge detection, image fusion

1. INTRODUCTION

A bone tumor, (also spelled bone tumour), is a

neoplastic growth of tissue in bone. Abnormal growths

found in the bone can be either benign (noncancerous) or

malignant (cancerous). Bone tumors may be classified as

"primary tumors", which originate in bone or from bone-

derived cells and tissues, and "secondary tumors" which

originate in other sites and spread (metastasize) to the

skeleton. Carcinomas of the prostate, breasts, lungs, thyroid

& kidneys are the primary carcinomas that most commonly

metastasize to the bone. Secondary malignant bone tumors

are estimated to be 50 to 100 times as common as primary

bone cancers. Primary tumors of bone can be divided into

benign tumors and cancers. Common benign bone tumors

may be neoplastic, developmental, traumatic, infectious, or

inflammatory in etiology. Some benign tumors are not true

neoplasms, but rather, represent hamartomas, namely the

osteochondroma. The most common locations for many

primary tumors, both benign and malignant include the

distal femur and proximal tibia. Since, by definition, benign

bone tumors do not metastasize, all secondary bone tumors

are metastatic lesions which have spread from other organs,

most commonly carcinomas of the breast, lung, and prostate.

Reliable and valid statistics on the incidence, prevalence,

and mortality of malignant bone tumors are difficult to come

by, particularly in the oldest (those over 75 years of age),

because carcinomas that are widely metastatic to bone are

rarely ever curable, biopsies to determine the origin of the

tumor in cases like this are rarely done.

The survival of bone cancer patients is related to

the extent of their disease at the time of diagnosis. In the

absence of distant metastases, the spread of tumors to the

mediastinal lymph nodes is a major determinant of both the

prognosis and the therapeutic approach. Proper staging is

important for selecting patients who may benefit from

surgical resection and for defining the treatment modalities

of patients who will undergo radiotherapy.

Tomographic imaging in nuclear medicine is based

on the metabolic activity of tissues and may be useful for

identifying pathological changes before they are detected by

radiological examinations such as CTs. CT provides the

anatomical structure and SPECT and PET and PET[11]

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2013

provides the functional information of organs and tissues.

PET (positron emission tomography) scans with 18F-FDG

(18F-fluorodeoxyglucose) have superior sensitivity and

specificity compared to chest CTs and are considered the

most accurate imaging method for staging patients with

bone cancers. However , there are limitations related to

positive predictive value of this method because there may

be FDG uptake in inflammatory cells. The sensitivity can

also be decreased when lymph node metastasis is

microscopic or below the spatial resolution threshold of

current, state-of-the-art scanners. The availability of PET is

restricted to a few institutions because of equipment costs.

Alternatively, single-photon emission computed tomography

(SPECT) is widely available, has lower costs than PET, and

does not require the presence of a cyclotron adjacent to the

hospital.

The association of functional images and

anatomical information from CTs may be useful in

interpreting SPECT by providing more accurate data

regarding the location and extent of tumor lesions. Hybrid

devices usually have dual detectors, with scintillation

cameras and low-dose CT scanners. Sequentially, data from

both CT and SPECT are acquired. The two images are

merged, creating SPECT images that are superimposed on

corresponding anatomical planes. This image fusion may

help to differentiate between tumors and other areas of

physiological activity.

Image fusion is the process of integrating

information from two or more images of the same position

into a single image that contains more information and is

more appropriate for visual perception. For the purpose of

most clinical applications, medical image fusion intends to

reduce ambiguity and minimize redundancy in fused image

while maximizing the relative information specifics.

This paper proposes a new method to detect

cancerous parts in body by using different image processing

techniques. Preprocessing techniques are done on SPECT,

CT images and then the edge of CT image is detected by

using a new method to enhance the disease affected part.

Then the SPECT and edge detected CT images are fused

with HIS,PCA and integrated retina inspired model (RIM)[2].

2. THE IHS, PCA AND RETINA INSPIRED

FUSION MODELS

2.1. The RGB-IHS Conversion Model

The IHS[3] transformation converts a multispectral image or

panchromatic image with red, green and blue channels

(RGB) to intensity, hue and saturation independent

components. The intensity displays the brightness in a

spectrum, the hue is the property of the spectral wavelength,

and the saturation is the purity of the spectrum. This

technique may be used for the fusion of multi-sensor

images.

To understand the whole fusion process preferably,

we must review the RGB-IHS conversion model. There are

two essential RGB-IHS conversion models. In this study, we

select a more close to the real visual effect model- triangular

spectral model. The IHS triangular model can produce a

fused and enhanced spectral image.

2.2. PCA Transform Fusion Approach

The whole idea of the method is described in detail

in References , and here the fundamentals of PCA[5] fusion

are briefly outlined as follows. Firstly, a multispectral image

is transformed with PCA transform and the eigen values and

corresponding eigenvectors of correlation matrix between

images in the multi-spectral image‟s individual bands are

worked out to obtain each matrix‟s principle components.

Next, the panchromatic image is matched by the first

principle component using histogram method. Finally, the

first principle component of the multispectral image is

replaced with the matched panchromatic image and with

other principle components, followed by the transformation

with inverse PCA transform to form the fused image.

2.3. Retina-Inspired Model

The RIM[6] fusion consists of five basic layers,

fusion structure diagram of which is depicted in Fig. 1. The

earliest layer represents an array of high resolution cone

photoreceptors, while the second layer is a high scale spatial

feature extractor. The third layer is the array of low

resolution receptors (horizontal cells), the fourth and the last

layers are made of bipolar and ganglion cells. Every layer

has its own mathematical model and corresponding

expressions.

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2014

fig1.Retina inspired model

In the proposed fusion approach, we use a black

box to represent the RIM model, which is just a part in the

whole system, its internal structure and some cells detailed

information in the fusion process shown the same as Fig. 1.

3. The Sobel edge detector

An edge[8], in an image, is a collection of connected

high frequency points. Visually, an edge is a region in an

image where there is a sharp change in intensity of the

image. Edge detection refers to the operation(s) performed

on an image to detect the edges in an image. The output of

edge detection is usually thresholded to retain only the edge.

Edge detection plays a vital role in object detection and

feature extraction and plays pivotal role in machine vision.

There are different types of edges – step edges, roof edges,

line edges, colour edges, gray level edges, texture edges etc.

Not all edges are detected by all edge detection operators.

Each operation has its specific specialty in edges and better

the edge detection, usually, more complex and costly is the

operation. An edge has both magnitude and direction. The

direction is used to identify the next possible edge point.

Finally, all the edge points are linked together to form an

object boundary. In an ideal scenario, all edges are

connected and we get a perfect boundary, but this is seldom

the case due to various reasons such as difference in

illumination, partially visible objects, texture variation,

inefficiency of the edge detection operator etc. Some times,

the edge detection operation might even result in false

edges. To resolve these issues, there have been a number of

edge detection operators defined throughout the history of

Image Processing. All the edge detection methods are based

upon the description of „changes of continuous functions

using derivatives‟ given by calculus. Simply put, all

operators are gradient operators.

Gradient edge detectors are those which describe

edges by means of partial derivatives. A change in the image

function can be described by a gradient that points in the

direction of the largest growth of the image function. They

are of three types: Operators performing derivatives by

using differences. Eg: Sobel , Operators based on the zero

crossings of the second derivatives. Eg: Canny edge

detector. , Operators that match the image function to a

known parametric model of the edges.

The well-known and earlier Sobel edge detector

computes the 2-D spatial gradient measurement on an

image. It consists of a pair of 3×3 convolution kernel:

Figure2: The convolution kernel for the Sobel Edge

detector. Note the emphasis on the horizontal and

vertical edges

The detector responds maximally to vertical and horizontal

edges, as can be observed from the values of the masking

kernel. It gives gradient nagnitude in both horizontal and

vertical direction.

The gradient component of Sobel edge operator:

( ) ( )

( ) ( )

4. PROPOSED METHOD

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2015

Figure3: proposed method

4.1 PREPROCESSING STEPS

Contrast Enhancement: Computed tomography (CT)

imaging is one of the areas that have seriously concerned

lately [1]. CT medical images are affected by different types

of degradations such as noise [2], blur [3] and contrast

imperfections. This method[9] handles the issue of enhancing

the contrast of CT medical images only. Generally, the

computed tomography (CT) medical images own a low

contrast, and enhancing the intensity of CT images is a vital

issue for humans. The contrast of any image is decided by

its dynamic range, which is defined as the ratio between the

brightest and the darkest pixel intensities. Many reasons led

to imperfection in the contrast such as varying display

devices, acquisition methods, transmission storage,

restoration and enhancement algorithms. Improving the

contrast of medical images would lead to better visual

representation of the images, improved syndromes

diagnosis, more accurate detection of diseases and as a

consequence enhancing the contrast of the CT images will

show smaller components and features of the medical image

allowing the therapist to perform a healthier treatment for

the patient. Enhancing the contrast of CT images is

considered as a pre-processing step in many medical image

processing applications.

When enhancing the contrast of computed

tomography (CT) medical images, two factors to complete

the task must be considered, those are speed and efficiency.

The proposed technique considers these two factors by

supplying a fast processing with effective results. This

technique has been utilized in the spatial domain. Moreover,

it has been applied to the entire image directly instead of

processing the image pixel by pixel. Normalizing the image

based on its size takes the following course: first, the size of

the processed image is determined. Then the enhancement

variable (K) is computed using the subsequent equation:

∑ ∑ ( )

Where (x) is the degraded image, the above

equation sums all pixels values and divides them by the size

of the image represented by (m) and (n). Finally, the image

is enhanced using the following equation:

( )

( ) ( )

Where (x) is the degraded image, (min, max) are the

minimum and maximum pixel values of the processed

image, (K) is the enhancement variable, and (EI) is the

contrast improved image. The following flow chart

represents the enhancement methodology:

Figure4 :Contrast Enhancement Method

Image Registration: Image registration[12] is the process of

transforming different sets of data into one coordinate

system. Data may be multiple photographs, data from

different sensors, from different times, or from different

viewpoints.[1] It is used in computer vision, medical

imaging, military automatic target recognition, and

compiling and analyzing images and data from satellites.

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2016

Registration is necessary in order to be able to compare or

integrate the data obtained from these different

measurements. Image registration essentially consists of

following steps Feature detection: Salient and distinctive

objects (closed-boundary regions, edges, contours, line

intersections, corners, etc) in both reference and sensed

images are detected.Feature matching: The correspondence

between the features in the reference and sensed image

established.Transform model estimation: The type and

parameters of the so-called mapping functions, aligning the

sensed image with the reference image, are estimated.Image

resampling and transformation: The sensed image is

transformed by means of the mapping functions.

Image resampling (RS): RS[12] is the procedure that

creates a new version of the original image with a different

width and height in pixels. Simply speaking, RS can change

the size of the image. Increasing the size is called

upsampling, for example. On the contrast, decreasing the

size is called downsamplig. Note that the spatial resolution

would not change after the RS procedure, either upsampling

or downsamplig.

Image normalization: A common problem associated with

the use of multisource image data is the grey value

differences caused by non-surface factors such as different

illumination, or sensor conditions. Such differences make it

difficult to compare images using same color metric system.

Image normalization[12] is required to reduce the radiometric

influences caused by non-surface factors and to ensure that

the grey value differences between temporal images reflect

actual changes.

4.2 EDGE DETECTION

Edge detection[1] is one of the fundamental steps in

image processing, image analysis, image pattern recognition

etc. Edge can outline the target objects profile, which is an

important attribute of extract from the images recognition.

Edge information in CT image is an important characteristic

because lung region contains an abundance of edges like

vessels, air tree, air sacs, artery branches etc.. In real world

applications, medical images contain object boundaries and

object shadows and noise. Therefore, they may be difficult

to distinguish

the exact edge from noise or trivial geometric features.

Fig5: Proposed edge detection

The proposed edge detection method can visualize

the disease affected part well and which is shown

below.Edge detection steps include :The input image is first

convoluted with the horizhontal and vertical components of

Sobel operator and second derivative of Sobel operator.

Then take the distance function between the outputs of

Sobel and Second derivative since second derivative

operator enhances only smaller edges and sobel operator

enhances all types of edges. Then taking the convolution of

horizontal and vertical components. By taking the edge

detection

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2017

Figure6 : edge detected images of bone

4.3 IMAGE FUSION

Fig. 7 shows the whole fusion process[2] of the

proposed approach, which may be divided into the following

several steps. Firstly, the MS image is transformed into the

IHS triangular model components. Then, histogram

matching is applied to match the histogram of the CT image

with the SPECT intensity component. Next, the PCA

transform extracts their own principal components of MS

intensity image and new CT (called New Pan), and selects

corresponding components‟ weight coefficients by

calculating their spatial frequency respectively to obtain a

new intensity component. Finally, the approach is performed

by combining the new intensity component and original

SPECT intensity component, using retina-inspired fusion

model.

In this stage a final intensity image is obtained,

which contains the same spatial detail of the original CT and

has the same intensity distribution to the original SPECT. In

the meantime, it also avoids some superfluous details and

artifacts in the previous transformation. Ultimately, we can

obtain a satisfied fused image by inverse IHS transform

exploiting the new intensity component and original H and S

components of SPECT image.

This fusion process generates a new high resolution

color image. The new image contains both the spatial details

of the CT source image and the spectral information of the

SPECT source image, simultaneously. How to select two

principal components‟ weight coefficients after PCA

transform is a critical problem. This paper propose an

adaptive selection method by calculating spatial frequency

(SF) of original MS intensity image and old CT image.

The expression for a K *L pixels image f (x, y) is defined as:

√

where RF and CF are the row frequency and column

frequency respectively.

√

∑ ∑ ( ) ( )

√

∑ ∑ ( ) ( )

Nisthula P, IJECS Volume2 Issue 6 June, year 2013 Page No.2012-2018 Page 2018

Fig7. Proposed fusion method

The selection of two principal components‟ weight

coefficients based on SF can be depicted as:

Where , represent the principal component of New Pan

and original multispectral intensity component,

respectively. and are normalized SF values.

Figure8: SPECT, CT images and SPECTCT fused

images of a patient with bone cancer.

5. CONCLUSION

This paper proposed a simple and easy image

fusion method to detect the bone cancer. Edge detected CT

is fused with the SPECT image to enhance visualization of

disease affected region. SPECT are shown in pseudo-color.

The SPECT produces images with suitable color and low

spatial resolution, while CT provides appropriate spatial

resolution with no color information content. This study

integrated the merits both preserving spatial information of

the IHS transform, reduces the spectral distortion by using

RIM and minimizing redundancy of PCA transformation

and obtained the satisfying fused results.

REFERENCES

[1] MA,Jing, BI,Qiang, “Processing practise of remote sensing

image based on spatial modeler‟‟ The National Natural Science

Foundation of China ( No. 41071237 ) IEEE 2012

[2] Changtao He, Guiqun Cao, Fangnian Lang, “An Efficient

Fusion Approach for Multispectral and Panchromatic Medical

Imaging”,Biomedical Engineering Research, March. 2013, Vol. 2

Iss. 1,.

[3] S. Daneshvar and H. Ghassemian, “MRI and PET image fusion

by combining IHS and retina-inspired models,” Information

Fusion, vol. 11, pp. 114-123, 2010.

[4] K. G. Baum, M. Helguera and A. Krol, “Fusion viewer: a new

tool for fusion and visualization of multimodal medical data

sets,”Journal of Digital Imaging, vol. 21, pp. 59–68, 2008.

[5] C. Wen, L. Bicheng and Z. yong, “A remote sensing image

fusion method based on PCA transform and wavelet packet

transform,” In: IEEE Conf. on Neural Networks &Signal

Processing, vol. 976-980, 2000.

[6] Sabalan Daneshvar, member, IAENG, Mosleh Elyasi, Morad

Danishvar, “Multispectral and Panchromatic Images Fusion Based

on Integrating Feedback Retina and IHS Model”, Proceedings of

the World Congress on Engineering 2011 Vol II WCE 2011, July 6

- 8, 2011, London, U.K.

[7] T. te-ming, S.shun-chi and S.Hsuen-chyunet, “A new look at

IHS-like image fusion methods,” Information fusion, vol. 2, pp.

177-186, 2001

[8] Om Pavithra Bonam and Sridhar Godavarthy “edge detection”

Computer Vision (CAP 6415 : Project 2)

[9] Zohair Al-Ameen, Ghazali Sulong and Md. Gapar Md. Johar “

Enhancing the Contrast of CT Medical Images by Employing a

Novel Image Size Dependent Normalization Technique”

International Journal of Bio-Science and Bio-Technology Vol. 4,

No. 3, September, 2012

[10] Disha Sharma, Gagandeep Jindal “Identifying Lung Cancer

Using Image Processing Techniques” International Conference on

Computational Techniques and Artificial Intelligence

(ICCTAI'2011)

[11] Heather A. Jacene, Sibyll Goetze, Heena Patel, Richard L.

Wahl and Harvey A. Ziessman, “Advantages of Hybrid SPECT/CT

vs SPECT Alone”, The Open Medical Imaging Journal, 2008, 2,

67-79

[12] Gang hong,Image fusion, image registration, and Radiometric

normalization for high resolution image processing, department of

geodesy and geomatics engineering university of new Brunswick